Selection of mature T cells in the thymus is dependent on interactions between the T cell receptor (TCR) and an MHC:self-peptide complex. Mature T cells are exported to the periphery where they have the potential to be activated by some foreign antigenic peptide in association with MHC. What is the relationship between the MHC-associated self-peptide that selects a mature T cell in the thymus and the foreign peptide that induces activation of the same T cell in the periphery? Since both peptides must, in association with MHC, interact with the same TCR they clearly cannot be randomly related. A number of observations suggest that in addition to epitope specificity TCR also recognizes other features of an MHC:peptide complex that are dependent on the interaction between peptide and MHC. These features may include MHC conformational determinants induced by bound peptide or a special topographical relationship between the peptide epitope and some MHC determinant that is also recognized by a specific TCR. In principle, it should be possible to define a class of peptides that preserve the same critical features of MHC:peptide interaction even though they express distinct epitopes. We suggest that complexes of MHC with many such peptides might interact with the same TCR albeit with altered affinity. Reduced affinity of interaction could result in positive rather than negative selection in the thymus and could give rise to partial agonist or antagonist activity in the periphery. We have used structural data available for specific MHC:peptide complexes to design a panel of such peptide analogs. We propose to test these peptides in an assay we have developed for peptide dependent positive selection of antigen-specific thymic precursors. Briefly, fetal thymus lobes are exposed to various peptides in organ culture and are subsequently grafted into SCID mice. Such grafts lead to functional reconstitution of SCID by donor derived T cells. It can then be determined by precursor frequency analysis whether a particular peptide analog enhances selection of mature T cells of defined specificity. A key feature of this assay is that positive selection is analyzed in thymus expressing a normal repertoire of diverse TCR. This makes it possible to address some issues that cannot be clearly addressed in a TCR transgenic model. Preliminary data is presented to demonstrate positive selection of antigen-specific thymic precursors by a specific peptide analog.